This may be the week the social media era's legal reckoning officially began. Juries in two separate courts found Meta and YouTube legally responsible for social media addiction and mental-health harm to minors. A California jury determined both companies were negligent in a case brought by a young woman who alleged the platforms served her harmful content from a young age.

The combined damages of $381 million barely register on Meta's balance sheet. What registers is the precedent: the first time major platforms have been found liable — not accused, found liable by juries — for the addictive design of their products. Thousands of similar cases are waiting in courts across the country.

Meta's response has been telling. The same week as the verdict, the company laid off roughly 700 employees across Reality Labs, recruiting, sales, and Facebook teams. But this isn't retreat: Meta guided to up to $135 billion in AI capital expenditure for 2026 and revealed four generations of custom AI chips — MTIA 300 through 500 — to deploy by the end of 2027.

He's not alone in that bet. Atlassian laid off 1,600 employees — 10% of its workforce — and replaced its CTO with two AI-focused CTOs. Apple is rolling out a reimagined, AI-powered Siri with iOS 26.4. The AI Accountability Act passed, requiring bias audits for AI used in hiring, lending, healthcare, and criminal justice.

And the models keep coming. March 2026 has produced more major releases than most entire quarters of 2024: GPT-5.4 with a 1.05 million token context window, Qwen 3.5 with native multimodal support, at least 12 major models and tools in the first week alone. The pace is accelerating, and the regulatory infrastructure is struggling to keep up.